Journal of Neurophysiology

Dopaminergic Modulation of Excitatory Postsynaptic Currents in Rat Neostriatal Neurons

Masashi Umemiya, Lynn A. Raymond


Umemiya, Masashi and Lynn A. Raymond. Dopaminergic modulation of excitatory postsynaptic currents in rat neostriatal neurons. J. Neurophysiol. 78: 1248–1255, 1997. γ-aminobutyric acid (GABA)-containing medium spiny neurons constitute ∼90% of the neuronal population in the neostriatum (caudate and putamen) and play an important role in motor programming. Cortical glutamatergic afferents provide the main excitatory drive for these neurons, whereas nigral dopaminergic neurons play a crucial role in regulating their activity. To further investigate the mechanisms underlying the dopaminergic modulation of medium spiny neuronal activity, we tested the effect of dopamine receptor agonists on excitatory synaptic transmission recorded from these neurons. Excitatory postsynaptic currents (EPSCs) were evoked by local stimulation and recorded from medium spiny neurons in postnatal rat striatal thin brain slices. Recordings were made using the whole cell patch-clamp technique under voltage clamp and conditions that selected for the α-amino-3-hydroxy-5-methyl-4-isoxazole propionate- and kainate-type glutamate receptor-mediated components of the EPSC. Incubation of slices in 10 μM dopamine resulted in a 33 ± 11% (mean ± SE) decrease in the amplitude of evoked EPSCs, an effect that developed during seconds. The relative variability in amplitude of dopamine's effects on medium spiny neuron EPSCs may reflect activation of different receptor subtypes with opposing effects. In contrast to the results with dopamine, incubation of slices in SKF 38393, a D1-type dopamine receptor selective agonist, resulted in dose-dependent potentiation of the medium spiny neuron EPSC that developed during several minutes. At a concentration of 5 μM, SKF 38393 resulted in a 29 ± 4.5% increase in EPSC amplitude, an effect that was blocked by preincubation with the D1-selective antagonist, SCH 23390 (10 μM). On the other hand, 5 μM SKF 38393 had no apparent effect on medium spiny neuron currents activated by exogenous application of glutamate or kainate. However, because of the inherent limitations of rapid agonist perfusion in the brain slice preparation (caused by slow agonist diffusion and rapid glutamate receptor desensitization) and because of anatomic evidence that colocalizes D1 and glutamate receptors to medium spiny neuron dendrites, our results leave open the possibility that the effect of D1 receptor activation on the EPSC is mediated via modulation of postsynaptic glutamate receptor responsiveness. The significant potentiation by D1 receptor agonists of EPSC amplitude at the cortico-striatal medium spiny synapse that we observed, in part, may underlie the role of D1 receptors in facilitating medium spiny neuronal firing, with implications for understanding regulation of movement.


Dopamine is involved in the physiological, as well as pathological, regulation of the basal ganglia and plays a crucial role in motor programming (Graybiel 1990; Nestler 1995; Surmeier et al. 1993). One pathway by which dopamine may act is by modulating glutamatergic input from the cortex and thalamus onto neostriatal GABAergic medium spiny neurons (MSNs) (Di Chiara et al. 1994). In the neostriatum, glutamate-gated ion channels mediate fast excitatory synaptic transmission from cortical afferents to MSNs (Jiang and North 1991; Kita 1996), whereas dopamine released from substantia nigral afferents activates a family of G-protein–coupled receptors resulting in modulation of intracellular cyclic-AMP (cAMP) levels (Grandy and Civelli 1992).

The molecular cloning of ionotropic glutamate receptors has confirmed their division into three subtypes, based on sequence homology as well as physiological and pharmacological characteristics (Hollmann and Heinemann 1994; Nakanishi 1992; Seeburg 1993). These subtypes are named for their preferred agonists: N-methyl-d-aspartate (NMDA),α - amino - 3 - hydroxy - 5 - methyl - 4 - isoxazole  propionate(AMPA), and kainate. As well, a family of dopamine receptors has been cloned recently and divided into D1 (D1, D5) and D2 (D2, D3, D4) subtypes, based on their positive or negative coupling to adenylyl cyclase, respectively (Grandy and Civelli 1992; Seeman and Van Tol 1994; Sibley and Monsma 1992). Because dopamine receptors regulate intracellular cAMP levels and all three subtypes of mammalian glutamate receptors have been shown to be modulated functionally by the activation of cAMP-dependent protein kinase (PKA) (Greengard et al. 1991; Raman et al. 1996; Raymond et al. 1993; Roche et al. 1996; Wang et al. 1991, 1993), activation of dopamine receptors potentially can regulate glutamate receptor activity.

Recent anatomic studies have localized both ionotropic glutamate receptors and dopamine receptors to dendrites of neostriatal MSNs (Hersch et al. 1995; Martin et al. 1993; Smith and Bolam 1990), raising the possibility that dopamine receptor activation may regulate glutamate receptor function and thereby modulate MSN postsynaptic responsiveness. On the other hand, D2 receptors also can be found localized to axon terminals, although presynaptic D1 receptors are rare in the striatum (Hersch et al. 1995; Levey et al. 1993; Smiley et al. 1994) but may be found in the nucleus accumbens (Nicola et al. 1996). Indeed, recent evidence supports both pre- and postsynaptic modulation by dopamine of cortical glutamatergic-MSN neurotransmission. One study suggests that activation of D2 receptors on cortical glutamatergic terminals results in inhibition of excitatory postsynaptic potentials (EPSPs) recorded from neostriatal MSNs (Hsu et al. 1995). Other data suggest that D1 agonists potentiate MSN EPSPs by enhancing postsynaptic glutamate receptor responsiveness (Cepeda et al. 1993).

To further elucidate the mechanisms by which dopamine modulates excitatory synaptic transmission in the striatum, we examined the effect of dopamine receptor agonists on evoked excitatory postsynaptic currents (EPSCs) recorded from visualized MSNs under voltage clamp using the whole cell patch-clamp recording method and blocking NMDA receptors as well as any potassium conductance. Our main finding is that D1 receptor activation facilitates excitatory synaptic transmission mediated by AMPA/kainate-type glutamate receptors, apparently without altering their agonist sensitivity. Our results provide further insight into mechanisms underlying dopaminergic modulation of MSN activity and may enhance understanding of dopamine's role in motor control.


Brain slice preparation

Rat brain slices were prepared essentially as previously described (Edwards et al. 1989; Umemiya and Berger 1994), and before being killed, rats were maintained according to regulations of the Canadian Council on Animal Care. Briefly, coronal brain slices (200–300 μm in thickness) were prepared from Wistar rats (P7–P16) using a vibratome. The ionic composition of the solution used for preparation of slices was (in mM) 120 NaCl, 25 NaHCO3, 1 NaH2PO4, 3 KCl, 2 CaCl2, 1 MgCl2, and 10 glucose. MSNs were visualized using a water immersion objective lens (Axioskop, ×40) and were identified on the basis of their location, in caudate and putamen, as well as by the ovoid cell body of 8–14 μm (major axis). Cholinergic neurons, another neuronal type in the striatum, were distinguished easily because of their large size (>15 μm) and bipolar shape.


Recordings were made at room temperature from MSNs under voltage clamp, using the patch-clamp recording technique in the whole cell configuration (Hamill et al. 1981). The holding potential was −65 mV, unless otherwise indicated. EPSCs were evoked by electrical stimulation of afferent fibers using a glass pipette containing external solution (tip resistance ∼5 MΩ) that was placed tens of micrometers (toward the cortex) from the MSN recording electrode. We monitored the EPSC amplitude for ∼10 min before the addition of any drugs and rejected responses that showed >10% run-down. Under our recording conditions, the EPSC amplitude was usually stable for ≥30 min. The external solution contained (in mM) 140 NaCl, 3 KCl, 2 CaCl2, 1 MgCl2, 10 glucose, 0.01 bicuculline, and 10 N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES)/NaOH, pH 7.4, and was perfused at a rate of ≥2 ml/min. Recording pipette solution contained (in mM) 120 CsMeSO4, 4 NaCl, 10 ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid, 1 MgCl2, 3 MgATP, 0.3 GTP-Tris, 10 HEPES/CsOH, pH 7.2. Recordings were made in the presence of bicuculline methiodide (10 μM) to block γ-aminobutyric acid-A receptors. Currents were amplified by the Axopatch 200A (Axon Instruments), filtered at 1–2 kHz, and digitized at 5 kHz for whole cell recording and were filtered at 2 kHz and digitized at 10 kHz for excised patch recording. We used pClamp software (Axon Instruments) for data acquisition and analysis. All drugs were applied for ∼10 min unless otherwise indicated. To determine the effect of drug application on EPSC amplitude in each experiment, the mean baseline EPSC amplitude during a 5- to 10-min period immediately preceding drug application was measured. Then the mean EPSC amplitude measured during a 5-min period 10–12 min after the start of drug application was normalized to this baseline. Results are reported as means ± SE.


Cesium hydroxide was obtained from Aldrich. Dopamine, SKF38393, (R)-SCH 23390, 2-amino-5-phosphono-valeric acid (APV),and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) were from RBI. All other chemicals were from Sigma. CNQX was dissolved in dimethyl sulfoxide at a concentration of 0.4 mM. Dopamine, SKF 38393, and SCH 23390 stock solutions were prepared fresh just before use, and 0.1% ascorbic acid was added to prevent oxidation.


Characterization of EPSCs recorded from MSNs

EPSCs were evoked by electrical stimulation using a glass pipette as described in methods. Figure 1 shows EPSCs recorded at various holding potentials. In these experiments, the membrane potential was stepped to each new value for ∼1 min before evoking an EPSC. At holding potentials less than or equal to −30 mV, EPSCs were evoked with a latency of 3.3 ± 0.4 ms, 10–90% rise-time of 1.3 ± 0.7 ms, and decay time constant of 3.6 ± 1.6 ms (n = 23). In contrast, at positive potentials, current rise and decay times were significantly slower, suggesting that EPSCs had at least two components (Fig. 1 A). The fast kinetic component contributed significantly to EPSCs at all holding potentials and was characteristic of non-NMDA receptor-mediated currents. The slower component, on the other hand, showed block at negative potentials in the presence of 1 mM MgCl2 and was therefore typical of NMDA receptor-mediated currents (Mayer and Westbrook 1987; Spruston et al. 1995). Moreover, CNQX, a specific non-NMDA receptor antagonist (Honoré et al. 1988), blocked fast decaying EPSCs at negative potentials in a dose-dependent manner (Fig. 1 B; n = 6), whereas APV (40 μM), an NMDA receptor antagonist, blocked slow EPSCs at positive potentials (Fig. 1 C; n = 4). Taken together, these results indicate that EPSCs evoked by stimulation of cortical afferents and recorded from striatal MSNs at negative holding potentials (in 1 mM MgCl2 and 10 μM bicuculline) were mediated almost exclusively by non-NMDA type glutamate receptors.

Fig. 1.

Excitatory postsynaptic currents (EPSCs) are mediated by both non-N-methyl-d-aspartate (NMDA) and NMDA-type glutamate receptors in medium spiny neurons (MSNs). Data shown in A–C are from 3 different representative cells. A: EPSCs exhibited a slower time course at depolarized potentials. EPSCs were evoked by single focal stimuli (16 V, 0.2 ms) at 0.5 Hz. Traces averaged from 20 trials at the indicated holding potentials are superimposed. B: 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; 10 μM) reversibly blocked EPSCs at negative potentials. EPSCs were evoked at 0.5 Hz. Left: each point represents the mean amplitude of EPSCs from 30 trials recorded at V h = −65 mV. Right: traces represent the average of 30 trials recorded at the times indicated (1–3) in the left panel. C: 2-amino-5-phosphono-valeric acid (APV; 40 μM) blocked EPSCs at positive potentials in the presence of CNQX (10 μM). EPSCs were evoked at 0.2 Hz, and each point shown in the left panel represents the mean amplitude of 10 trials, recorded at +50 mV. Each trace shown in the right panel is the average of 10 EPSCs recorded in the presence (2) and absence (1 and 3) of APV. Rats used for slice recordings shown in A–C were postnatal age 7 days.

Synaptic modulation by dopamine receptor agonists

We tested the effect of dopamine (10 μM) on EPSCs (Fig. 2 A). To selectively measure the effect of dopamine on non-NMDA receptor-mediated EPSCs, the membrane potential was held at −65 mV, and 1 mM MgCl2 was included in the external solution. Under these conditions, dopamine significantly suppressed mean EPSC amplitude by 33 ± 11% (n = 4 of 4 cells). The decrease in EPSC amplitude developed rapidly (within tens of seconds, see Fig. 2 A) and appeared to reverse just as rapidly.

Fig. 2.

Time course of the effect of dopamine agonists/antagonists on EPSCs. The effect of dopamine (A), SKF 38393 (B), and SCH 23390 + SKF 38393 (C) application on EPSCs recorded from 3 different representative MSNs are shown. A: traces (left) and points (right) are averages of 20 trials. EPSCs were evoked at 0.5 Hz, as described in Fig. 1. B: traces in the left panel and points in the right panel are averages of 10 trials. EPSCs were evoked at 0.2 Hz. C: Traces (left) and points (right) are averages of 18 trials. EPSCs were evoked at 0.2 Hz. Traces shown in A–C were recorded at the times indicated by numbers (1 and 2) shown on the plots. Postnatal ages of rats used for slice recording in A–C were 11, 13, and 9 days, respectively.

Presynaptic terminals of cortical afferent fibers are known to have D2 receptors (Di Chiara et al. 1994; Hersch et al. 1995), and activation of D2 receptors has been shown to suppress EPSPs (Hsu et al. 1995). Therefore the effects of D1 receptor activation could be masked by presynaptic inhibition caused by D2 receptor activation during application of dopamine. To examine the effects of D1 receptor activation alone, we used the selective D1 agonist SKF 38393 (Seeman and Van Tol 1994). The addition of 5 μM SKF 38393 resulted in significant potentiation of the mean EPSC amplitude by 29 ± 4.5% (n = 5 of 5 cells; Figs. 2 B and 3). In contrast to dopamine's effect, potentiation by SKF 38393 developed slowly, during minutes (time to maximum potentiation was 4.0 ± 1.2 min, n = 5). The effect of SKF 38393 was dose dependent: incubation of slices in 1 μM SKF 38393 resulted in EPSC amplitudes that were not significantly different from baseline (Fig. 3; n = 6); whereas after treatment with 25 μM SKF 38393, EPSC amplitudes were potentiated by 34 ± 1.4% in three of five cells tested, but the two other cells showed no significant change from baseline (not shown). However, in contrast to potentiation by 5 μM SKF 38393, maximum potentiation of EPSC amplitude by 25 μM SKF 38393 occurred several minutes after wash-out of the drug. Finally, the D1-selective antagonist SCH 23390 (10 μM) completely blocked the effect of 5 μM SKF 38393 (Figs. 2 C and 3; n = 4). In subsequent experiments described below, we tested whether SKF 38393 affected the agonist-evoked current response of postsynaptic non-NMDA receptors to elucidate the site of synaptic modulation.

Fig. 3.

Pooled data showing effect of D1 receptor agonists/antagonists on EPSC amplitude. Normalized EPSC amplitude was calculated as described in methods. Bars represent means ± SE for n = 4–6 different cells. *, significant difference in normalized postsynaptic current amplitude between the 5 μM SKF 38393-treated group and the other 3 groups as determined by the analysis of variance, using the Fisher protected least significant difference PLSD test at the 95% confidence level (by this test, none of the other 3 groups were significantly different from 1 another). SKF, SKF 38393; SCH, SCH 23390. Postnatal ages of rats used for slice recordings ranged from P7 to P16.

Glutamate-evoked currents in MSNs

We activated postsynaptic glutamate receptors by pressure application of 10 mM glutamate in the presence of tetrodotoxin (1 μM). We used a small glass pipette (<1 μm) positioned ≥50 μm away from the soma to optimize activation of dendritic glutamate receptors (in some cases, dendrites were identified under the microscope), because dopamine receptors are distributed preferentially on distal dendrites of MSNs (Hersch et al. 1995). Usually, currents evoked by glutamate exhibited rapid activation and decay kinetics at a holding potential of −65 mV in Mg2+-containing external solution (Fig. 4). We accepted only rapidly activating currents, with a 10–90% rise time of <10 ms. Usually, glutamate-evoked currents were biphasic, and we only measured the peak amplitude of the fast response. In 11 cells tested, analysis of the fast component of the current response revealed a 10–90% rise time of 8.7 ± 1.4 ms, decay time constant of 18.9 ± 9.1 ms, and mean peak current amplitude of 30.4 ± 14.5 pA. The fast component showed a reversal potential ∼0 mV, and the current-voltage (I-V) relationship was almost linear (Fig. 4 A). The slow component also reversed at ∼0 mV, but the I-V curve showed outward rectification (Fig. 4 A). Pharmacologically, the fast component of the glutamate response was abolished almost completely by 10 μM CNQX (peak amplitude was reduced by 78.2 ± 14.7%; n = 5), whereas block of the slow component was incomplete at this concentration of CNQX (Fig. 4 B). Together with the I-V relationship, these results suggest that the slow component was composed of NMDA receptor-mediated current, as well as a contribution from non-NMDA receptors activated by the slow diffusion of glutamate. Moreover, we conclude that the fast component was mediated by non-NMDA receptors on dendritic sites located near the application pipette.

Fig. 4.

Characterization of glutamate-activated current recorded from MSNs. A: I-V relationship of glutamate-activated currents. From a holding potential of −60 mV, voltage was stepped in 15-mV increments to potentials ranging from −60 to +60 mV. Each potential was held for ∼1 min before applying glutamate (10 mM) by pressure pulse (150 kPa, 20 ms) from a pipette placed 50 μm from the cell soma. Left: peak current amplitude (♦) and current amplitude measured 100 ms after the initiation of the glutamate pulse (×) were plotted as a function of holding potential. Right: corresponding current traces; ▿ and ▾, time points at which the peak and “steady state” current amplitudes were measured, respectively. B: CNQX (10 μM) sensitivity of glutamate-activated currents at a holding potential of −65 mV. The amplitude of the fast component of the glutamate (10 mM)-evoked current responses was measured at a time point corresponding to the peak response for the control current (i.e., before addition of CNQX) and plotted against time (left). Right: superimposed current traces shown recorded in the presence (2) and absence (1 = control, 3 = wash) of CNQX. Postnatal age: 8 days.

Previous studies have shown that cAMP-dependent PKA-mediated phosphorylation of non-NMDA receptors potentiates the peak current response to fast application of glutamate without affecting the amplitude of the steady-state (nondesensitizing) component (Greengard et al. 1991; Raymond et al. 1993). To assess whether the peak glutamate-evoked current response was attenuated by receptor desensitization due to slow agonist application, we recorded currents from outside-out patches excised from the soma of MSNs and measured the time course of non-NMDA receptor activation and desensitization (Fig. 5). We used the same method described previously for glutamate application to MSN dendrites while in the whole cell recording mode, except that for these recordings in the outside-out patch mode, the application pipette was placed only a few micrometers from the recording pipette. Again, to record non-NMDA receptor-mediated currents, the external solution contained 1 mM MgCl2 without added glycine. The reversal potential of the glutamate-activated currents was close to 0 mV (Fig. 5 A;n = 4), and currents were sensitive to CNQX (10 μM;n = 4; not shown). Although the mean rise time (3.1 ± 1.1 ms) for glutamate-activated current was faster than that seen with dendritic application, the decay time constant was similar to that of dendritic application (14.7 ± 8.1 ms; n = 8). Furthermore, the current decay time constant was independent of the duration of pressure application of glutamate (Fig. 5 B), suggesting that the time course for current decay was determined largely by receptor desensitization rather than by glutamate dissociation and diffusion (cf. Spruston et al. 1995). Because the rise time of whole cell glutamate-activated currents was less than twofold faster than the mean decay time constant of current responses recorded from excised patches, it is possible that the peak amplitude of whole cell glutamate-evoked currents was significantly attenuated by receptor desensitization.

Fig. 5.

Time course of desensitization of glutamate-evoked non-NMDA receptor-mediated current responses recorded in excised patches from MSNs. A: glutamate (10 mM) was applied by pressure pulse (30 kPa, 5 ms) from a pipette placed ∼5 μm from the recording pipette. External solution contained 1 mM MgCl2 and no added glycine. Current traces (left) and points (right) represent the average of 5 trials at each test voltage. B: current responses activated by glutamate (10 mM) pressure pulses (200 kPa) of 20- and 100-ms duration are shown superimposed. Traces are the average of 5 trials. Postnatal ages were 14 and 7 days for A and B, respectively.

SKF 38393 shows no effect on glutamate-evoked currents recorded from MSNs

We tested the effect of SKF 38393 on the amplitude of the current response to glutamate and kainate in whole cell recordings from MSNs (Fig. 6). For ∼10 min before the addition of SKF 38393, we monitored the current responses to glutamate or kainate and rejected neurons showing >10% run-down. Application of 5 μM SKF 38393 for ≤5 min resulted in no significant change in the peak amplitude of the current response to glutamate (Fig. 6 A; n = 8). Because previous studies have shown a potentiating effect of PKA and/or dopamine on currents evoked by kainate (Knapp and Dowling 1987; Liman et al. 1989; Smith et al. 1995; Wang et al. 1991), which activates AMPA receptors in a nondesensitizing manner, we also tested the effect of SKF 38393 on kainate (10 mM)-activated currents. In contrast to glutamate, kainate was applied to both dendrites and soma. As expected, the activation and decay of kainate-evoked currents was significantly slower than that of glutamate current responses with a mean rise-time of 85.5 ± 15.4 ms, and the mean amplitude of the kainate response was 118.8 ± 26.0 pA (n = 5). Still, SKF 38393 (5 μM) was without effect on kainate-evoked currents (n = 5; Fig. 6 B).

Fig. 6.

Lack of effect of SKF 38393 (5 μM) on whole cell current responses evoked by glutamate agonists and recorded from 2 different representative MSNs. A: effect of SKF 38393 on glutamate (10 mM)-activated currents. Current responses to glutamate application (pressure pulse of 200 kPa, 20 ms) at a frequency of 0.1 Hz were recorded at V h = −65 mV. Representative current traces before and during application of SKF 38393 are shown (left) as is peak amplitude of the glutamate-evoked current response plotted vs. time (right). B: effect of SKF 38393 on kainate (10 mM)-activated currents. Kainate was applied at 0.1 Hz, and current responses were recorded at V h = −65 mV. As in A, representative current traces are shown at the left, and current amplitude is plotted as a function of time in the right panel. Postnatal ages were 9 and 13 days for A and B, respectively.


We have found that activation of the D1 subtype of dopamine receptors by SKF 38393 potentiates excitatory synaptic transmission between cortical afferents and striatal MSNs in brain slices from early postnatal rats, whereas application of dopamine, which activates both D1 and D2 receptors, results in a more variable decrease in EPSC amplitude. The latter result is basically in agreement with previous work showing suppression of MSN EPSPs by dopamine and D2 agonists (Hsu et al. 1995). We also have shown that under conditions used for our recordings (i.e., voltage clamp at −65 mV in 1 mM Mg2+ and 10 μM bicuculline), short-latency evoked EPSCs are mediated predominantly by non-NMDA type glutamate receptors. Although we did not include acetylcholine receptor (AChR) antagonists in our recording solution, a recent study has shown that these receptors do not contribute to EPSPs evoked by intrastriatal stimulation at low (<1 Hz) frequency (Kita 1996). Moreover, EPSCs mediated by nicotinic AChRs would be disynaptic and thus would be expected to show longer latency than the responses analyzed in our study.

In our experiments, apparently opposing effects of dopamine and SKF 38393 were manifested during different timecourses—seconds versus minutes, respectively—suggesting perhaps that these two drugs ultimately affected different intracellular pathways. In this regard, it is interesting to note that the initial rapid decrease in EPSC amplitude observed with dopamine application appeared to be reversed just as rapidly (in some cases, with an overshoot—see Fig. 2 A); perhaps this “reversal” was in part due to the delayed effect of D1 receptor activation. In addition, incubation with a high concentration of SKF 38393 (25 μM) resulted in either no potentiation (2 of 5 cells) or else the development of maximum potentiation several minutes after washout of the drug (3 of 5 cells). Previous studies have demonstrated that the EC50 of SKF 38393 for increasing neostriatal intracellular cAMP levels is on the order of 0.4 μM (Andersen and Jansen 1990), suggesting that 5 μM SKF 38393 should be saturating for effects mediated by D1-type dopamine receptors. Moreover, higher concentrations of SKF 38393 may activate D2-type dopamine receptors (Andersen and Jansen 1990; Seeman and Van Tol 1994). Thus we speculate that our results with 25 μM SKF 38393 may reflect coactivation of D1- and D2-type receptors during perfusion of the drug. Because the effect on EPSCs of D1 activation appears to be relatively long-lasting (Fig. 2 B), whereas that of coactivation of D1 and D2 by dopamine is reversed rapidly (Fig. 2 A), washout of 25 μM SKF 38393 may “unmask” the effect of D1 activation.

In contrast to observed effects on EPSC amplitude, we have found no apparent effect of SKF 38393 on non-NMDA receptor-mediated current responses evoked by exogenous application of glutamate or kainate to MSNs (but see discussion below), suggesting that SKF 38393 modulates synaptic transmission through activation of presynaptic D1 receptors. Previous studies in other brain regions have demonstrated effects of presynaptic D1 receptor activation on glutamatergic transmission: D1 agonists were shown to potentiate glutamate release in ventral tegmentum (Kalivas and Duffy 1995), whereas they inhibited glutamatergic transmission in nucleus accumbens (Pennartz et al. 1992), apparently via presynaptic receptors (Nicola et al. 1996). Although anatomic studies in adult rat neostriatum indicate that D1 receptors rarely are localized to axon terminals (Hersch et al. 1995), perhaps there are presynaptic D1 receptors transiently expressed during the early postnatal period, or else present at levels below detection by immunocytochemical methods, that are responsible for mediating the effects of SKF 38393 in our system. As an aside, we noted no correlation between magnitude of EPSC modulation by SKF 38393 and developmental age of the rats in the range studied (P7–P16). However, the number of cells tested at any particular developmental age were quite small, and further experiments would be required to investigate whether developmental stage plays a role in determining potentiation of non-NMDA receptor-mediated EPSCs by D1-type dopamine receptor activation.

Alternatively, activation of postsynaptic D1 receptors may indeed modulate postsynaptic responsiveness at cortico-striatal MSN synapses in our system, but we may be unable to detect this effect due to the limitations of rapid agonist perfusion in the brain slice. Electron microscopic studies by Hersch et al. (1995), showing that a majority of glutamatergic cortical axon terminals synapse on MSN dendritic spines containing D1 receptors, strongly suggest a postsynaptic locus for the effects of SKF 38393. Moreover, there are several plausible explanations for the observed lack of effect of SKF 38393 on MSN glutamate receptor responsiveness in our system. First, our data indicate that the rise-time for complete solution exchange with application of glutamate to MSN dendrites while recording in the whole cell mode is less than twofold faster than the decay time constant for desensitization of current evoked in outside-out patches. These results suggest that the peak glutamate-evoked current amplitude recorded in the whole cell mode may be attenuated by ≤45%. Because previous studies have shown that D1 agonists and/or PKA-dependent phosphorylation potentiate the peak glutamate and/or kainate-evoked current amplitude (Greengard et al. 1991; Liman et al. 1989; Raymond et al. 1993; Roche et al. 1996; Smith et al. 1995; Wang et al. 1991, 1993), it is very likely that such an effect on glutamate-activated currents would go undetected in our system. However, our data also show no effect of D1 agonists on kainate-evoked currents, further suggesting that AMPA-type receptors are not modulated by D1 agonists (but see below), because AMPA receptors do not desensitize in response to kainate (Hollmann and Heinemann 1994). In contrast, kainate-type receptors show rapid desensitization to bothkainate and glutamate (Hollmann and Heinemann 1994), and therefore a modulatory effect of D1 receptor activation on these receptors is still possible; notably, the kainate-type receptor GluR6 has been shown to be potentiated by PKA-mediated phosphorylation in a heterologous expression system (Raymond et al. 1993; Wang et al. 1993). Second, it may be that only synaptic glutamate receptors, which are localized nearby D1 receptors on dendritic spines (Hersch et al. 1995), are modulated by D1 receptor activation. Because exogenous application of glutamate agonists would activate both synaptic and extrasynaptic glutamate receptors (especially in the case of kainate, which was applied to dendrites and soma), the modulatory effects of D1 receptors on synaptic glutamate receptors may be difficult to detect. Finally, it is also possible that modulation of non-NMDA receptor function by D1 agonists is dependent on the concentration of glutamate agonist used to elicit a current response (Wang et al. 1993). With our method of pressure ejection for applying agonists in the brain slice (and the rapid uptake of glutamate by glia), the precise concentration of glutamate seen by dendritic receptors is unknown.

Our data, obtained by using the patch-clamp recording technique under voltage clamp and conditions that minimize voltage-gated conductances and the effect of intrinsic MSN cable properties, provide some of the most direct evidence for potentiation of glutamatergic transmission at the cortico-striatal MSN synapse by activation of D1 receptors. Previous studies have documented effects of dopamine, as well as D1 and D2 receptor agonists, on excitability of MSNs to injection of current, or else on glutamatergic EPSPs, under current clamp using intracellular recording techniques. Akaike et al. (1987) showed that the effects of dopamine on MSN excitability were concentration dependent and varied with the receptor subtype (D1 vs. D2) activated, whereas Calabresi et al. (1987) showed a similar effect of dopamine on excitability and also examined its effects on EPSPs; both groups concluded that D1 agonists mimicked the effects of micromolar concentrations of dopamine and inhibited excitability and/or suppressed EPSPs. Hsu et al. (1995) demonstrated a similar depression of EPSPs due to micromolar dopamine, but found that this effect was mimicked by D2 agonists. In contrast, Cepeda et al. (1993) concluded that D1 agonists potentiated the majority of NMDA- and AMPA-induced MSN depolarizations, whereas these were inhibited by D2 agonists. In these studies using the intracellular recording technique under current clamp, MSN excitability and EPSP amplitude depends, at least in part, on the resting input resistance, resting membrane potential, and the relative activity of ion channels activated by depolarization. Because dopamine already has been shown to regulate the activity of voltage-gated channels, including sodium channels (Schiffmann et al. 1995; Surmeier and Kitai 1993; Surmeier et al. 1993), calcium channels (Surmeier et al. 1995), and potassium channels (Greif et al. 1995; Surmeier and Kitai 1993), it can be difficult to determine from intracellular current-clamp recordings whether dopamine actually modulates glutamate receptor-mediated responsiveness in the postsynaptic membrane of MSNs.

Although we cannot conclude from our results that the D1 agonist SKF 38393 potentiated glutamate receptor responsiveness in our system, such an effect has been demonstrated previously in studies in a variety of in vitro systems, including teleost retinal cells (Knapp and Dowling 1987), the goldfish Mauthner cell (Pereda et al. 1994), and chick motoneurons (Smith et al. 1995). Furthermore, several studies have shown that activation of cAMP-dependent protein kinase (either by D1 receptor activation or by more direct biochemical methods of activating the kinase) can result in potentiation of non-NMDA receptor-mediated current responses in teleost retinal cells (Liman et al. 1989), hippocampal neurons (Greengard et al. 1991; Wang et al. 1991), the goldfish Mauthner cell (Pereda et al. 1994), and chick motoneurons (Smith et al. 1995) and for recombinant receptors in a heterologous expression system (Raymond et al. 1993; Roche et al. 1996; Wang et al. 1993). Moreover, a recent study using the intracellular current-clamp method to record from MSNs in adult rat neostriatal slices found that PKA activation resulted in potentiation of glutamatergic EPSPs, as well as of depolarizations induced by glutamate agonists (Colwell and Levine 1995). Taken together with our results, these studies strongly suggest that D1 receptor activation may potentiate glutamate receptor-mediated current responses in MSNs, through activation of the cAMP-PKA pathway.

A variety of studies have suggested that D1 receptor activation in the neostriatum results in stimulation of the direct pathway (MSNs projecting to substantia nigra pars reticulata and internal globus pallidum), thereby disinhibiting thalamic neurons and facilitating movement (for review, see Albin et al. 1989; Di Chiara et al. 1994; Wichmann and DeLong 1993). By enhancing excitatory postsynaptic currents mediated by non-NMDA receptors, which we and others have shown to be primarily responsible for mediating fast excitatory transmission at the cortico-striatal MSN synapse (at least at normal hyperpolarized resting membrane potentials), D1 receptor activation would increase the likelihood of triggering an action potential at the MSN soma. This effect, in part, may explain the D1 receptor-mediated potentiation of direct pathway MSN transmission and provide further insight into mechanisms involved in motor control.


We thank Drs. Tim Murphy, Peter Reiner, and Chris Price for useful discussions and comments on the manuscript. We also are grateful to S. Sturgeon for assistance in manuscript preparation.

This work was supported by the Parkinson Foundation of Canada and by Medical Research Council of Canada Grant MT-12699 to L. A. Raymond.


  • Address for reprint requests: L. A. Raymond, Division of Neurological Sciences, Dept. of Psychiatry, University of British Columbia, 2255 Wesbrook Mall, Vancouver, British Columbia V6T 1Z3, Canada.

  • Present address of M. Umemiya: Dept. of Neurophysiology, Tohoku University School of Medicine, Sendhai 980-77, Japan.


View Abstract